[WIP] Windows build #42

bilderbuchi · 2021-06-14T15:23:48Z

Checklist

Used a personal fork of the feedstock to propose changes
Bumped the build number (if the version is unchanged)
Re-rendered with the latest conda-smithy (Use the phrase @conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
Ensured the license file is being packaged.
Try to use - {{ compiler('m2w64_cxx') }} # [win] again and re-render the recipe afterwards!

Let's see how far we get in adding a Windows build of gdb.
See #6.

conda-forge-linter · 2021-06-14T15:23:52Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

bilderbuchi · 2021-06-14T18:44:46Z

@conda-forge-admin , please rerender

conda-forge-linter · 2021-06-15T09:10:57Z

Hi! This is the friendly automated conda-forge-linting service.

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

Selectors are suggested to take a <two spaces>#<one space>[<expression>] form. See lines [27]

conda-forge-linter · 2021-06-15T12:20:21Z

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

bilderbuchi · 2021-06-15T14:28:48Z

um, I'm getting up to the make stage now, but I get "configure: error: C compiler cannot create executables" (link)

No idea what to do about that, does anybody have input here?

Start bld.bat using snippets from AnacondaRecipes [1] The script calls build.sh to reduce duplication. Remove packages unavailable on Windows. Add necessary m2 packages. [1] https://github.com/AnacondaRecipes/conda-feedstock/blob/master/recipe/bld.bat

bilderbuchi · 2021-07-01T19:15:08Z

@conda-forge-admin, please rerender

…nda-forge-pinning 2021.07.01.17.09.14

bilderbuchi · 2021-07-01T19:57:41Z

Hm, this fails with

conda.exceptions.ResolvePackageNotFound: 
  - ncurses

although I specifically first excluded ncurses, then replaced it with m2-ncurses (both approaches fail).

I can't recognize where/why it needs ncurses. The log output before even seems to suggest that conda's finished, but then another solve fails?!

Curiously enough, this did not occur in the previous incarnation of this PR. Maybe a change in the recipe rerender template?

bilderbuchi · 2021-07-01T20:10:27Z

@klauer: interestingly, the re-render above (181521f) apparently broke (only) the macos py39 build, although no mac files seems to have been touched?

klauer · 2021-07-01T20:59:32Z

Strange - diffing the working build output to the non-working one, I don't see many differences:
This PR's build:

conda-forge-ci-setup-3.9.4
Test phase fails with the gdb output During startup program terminated with signal ?, Unknown signal. which happens before even a thread gets created (*)
Doesn't touch any of the travis config as you noted

(*) Test diff

Working:

Developer mode is already enabled.
YES (0)
[New Thread 0x1503 of process 40618]
[New Thread 0x2803 of process 40618]
warning: `/BuildRoot/Library/Caches/com.apple.xbs/Binaries/Libc_darwin/install/TempContent/Objects/Libc.build/libsystem_darwin.dylib.build/Objects-normal/x86_64/darwin_vers.o': can't open to read symbols: No such file or directory.
warning: `/BuildRoot/Library/Caches/com.apple.xbs/Binaries/Libc_darwin/install/TempContent/Objects/Libc.build/libsystem_darwin.dylib.build/Objects-normal/x86_64/dirstat.o': can't open to read symbols: No such file or directory.
warning: `/BuildRoot/Library/Caches/com.apple.xbs/Binaries/Libc_darwin/install/TempContent/Objects/Libc.build/libsystem_darwin.dylib.build/Objects-normal/x86_64/dirstat_collection.o': can't open to read symbols: No such file or directory.
warning: `/BuildRoot/Library/Caches/com.apple.xbs/Binaries/Libc_darwin/install/TempContent/Objects/Libc.build/libsystem_darwin.dylib.build/Objects-normal/x86_64/init.o': can't open to read symbols: No such file or directory.
warning: `/BuildRoot/Library/Caches/com.apple.xbs/Binaries/Libc_darwin/install/TempContent/Objects/Libc.build/libsystem_darwin.dylib.build/Objects-normal/x86_64/variant.o': can't open to read symbols: No such file or directory.
Hello, World!
[Inferior 1 (process 40618) exited normally]

Not working:

Developer mode is already enabled.
YES (0)
During startup program terminated with signal ?, Unknown signal.
[New Thread 0x1503 of process 40602]
Tests failed for gdb-10.2-py39h593ab31_0.tar.bz2 - moving package to /Users/travis/miniforge3/conda-bld/broken
WARNING:conda_build.build:Tests failed for gdb-10.2-py39h593ab31_0.tar.bz2 - moving package to /Users/travis/miniforge3/conda-bld/broken
WARNING conda_build.build:tests_failed(2955): Tests failed for gdb-10.2-py39h593ab31_0.tar.bz2 - moving package to /Users/travis/miniforge3/conda-bld/broken

Working build

conda-forge-ci-setup-3.9.5 (no significant differences as far as I can tell, recipe here)
Everything else looks the same

I'd imagine the tests aren't so flaky as the 10.2 PR merged and built without issue...

bilderbuchi · 2021-07-02T05:49:35Z

Curiously, on the next run, both macos-py36 and macos-py39 failed, although again no macos config was touched. Maybe something is flaky after all.

Btw, easter-egg time, gdb still contains python 2 code, although that does not fail the build 😁

bilderbuchi · 2021-07-02T05:54:59Z

Meanwhile, there's conda-forge-ci-setup 3.9.6 released, so
@conda-forge-admin, please rerender

github-actions · 2021-07-02T05:56:02Z

Hi! This is the friendly automated conda-forge-webservice.
I tried to rerender for you, but it looks like there was nothing to do.

github-actions · 2021-07-02T08:59:05Z

Hi! This is the friendly automated conda-forge-webservice.
I tried to rerender for you, but it looks like there was nothing to do.

github-actions · 2021-07-02T09:01:58Z

Hi! This is the friendly automated conda-forge-webservice.
I tried to rerender for you, but it looks like there was nothing to do.

phil-blain · 2021-07-03T15:56:20Z

I'll try to help if I can for the Windows build, but I have close to zero experience on Windows.

Regarding the macOS Travis run that seems flaky, upon re-reading the GDB bugzilla entry linked in recipe/README, I think that I would have to update the patch that we use for macOS. I'll try to do that some time. In the meantime I think we can ignore macOS failures in this here PR, at least the During startup program terminated with signal ?, Unknown signal. error.

Regarding the latest macOS py36 failure, I see this at the end:

Solving environment: ...working... failed
failed to get install actions, retrying.  exception was: Unsatisfiable dependencies for platform osx-64: {'gdb==10.2=py36h7a1bca3_1'}
WARNING:conda_build.build:failed to get install actions, retrying.  exception was: Unsatisfiable dependencies for platform osx-64: {'gdb==10.2=py36h7a1bca3_1'}
WARNING conda_build.build:test(2859): failed to get install actions, retrying.  exception was: Unsatisfiable dependencies for platform osx-64: {'gdb==10.2=py36h7a1bca3_1'}

I'm not sure about this one. I don't see anything wrong before that. There's a lot of "is not an ELF" lines but that might just be conda-build acting out, I'm not sure. It's the same thing for the macOS py39 failure in 181521f...

Avoids a warning on Windows build.

bilderbuchi · 2021-08-03T16:17:47Z

yeah, I think we both hit a temporary CI breakdown, which has since resolved. I managed to get over some more roadblocks, and checked out the changes of that PR, but at this point I think my problems have a different source (\ in windows paths tripping up this dwarf function), but it's hard to troubleshoot (and find the time+energy for that).

bilderbuchi · 2021-08-03T16:24:25Z

Specifically, I suspect I need to add one more env vars to the cygpath conversion done here, but I currently don't know how to find out which path is the culprit. I think the warnings about escapes surrounding the error linked above could each stem from the initial letters of directories in the affected path, but I have not yet tried to decipher that.

bilderbuchi · 2021-08-20T22:11:04Z

The DEBUGDIR path in the lines 1 and 2 (and what follows in the same line is suspiciously similar to the pattern of invalid escape sequence warnings.
I suspect that this somehow (how?!) mangles forward-slash delimited paths as in the command line invocation into backslash delimited paths (with some expansion as a capital \L is also in there?), which then trigger all these invalid escape sequence warnings?

It's probably that --with-separate-debug-dir flag, I'll have to find out how to set that correctly on Windows.

bilderbuchi · 2021-08-21T11:52:18Z

OK, I disabled that flag on Windows, now it fails because it does not find netdb.h, which is apparently not included in mingw-64:

D:/bld/gdb_1629543547423/work/gdbsupport/netstuff.cc:28:19: fatal error: netdb.h: No such file or directory

Curiously enough, above that point it checks for the presence of netdb.h and correctly recognizes that it's not available, so I don't know why it stumbles over that later.

I looked at the various patches linked above but noticed nothing that seemed applicable.

bilderbuchi · 2021-08-24T19:34:22Z

Since gdb is not available from conda-forge for Windows, and it seems to me it should be, I have been trying to get that to build.
Apparently, I don't know what I'm doing, because I've been banging my head against this over the summer. At this point I'm (again) stuck.

Maybe some Windows/m2w64/mingw/gcc experts can take a look and help out getting gdb built? @conda-forge/help-c-cpp @conda-forge/core

bilderbuchi · 2021-08-24T19:45:07Z

My reasoning for going the route of a slim bld.bat shim that calls build.sh was to minimize the additional maintenance effort for the Windows build by not introducing another build script. I saw that approach in another package, so it seemed reasonable, but maybe it isn't?

jaimergp · 2021-08-25T07:57:03Z

Looks like a missing header: netdb.h. This seems to be not part of MingW, so you'll need to patch it to replace it by a Windows equivalent. Source 1, source 2, source 3.

Edit: oops sorry didn't read the full conversation. You might have tried this already 😬 . Anyway, you can see a similar patch for the MingW Python here. Worth a try? I don't know why they are not using it on their GDB build though.

Edit 2: It looks like configure should handle this but it does not. The offending file will check for the variable USE_WIN32API to use the Windows API or the Unix one, so I guess that one is not being set appropriately. Can you force define it in the calling script?

jaimergp · 2021-08-25T08:00:26Z

recipe/meta.yaml

-    - make >=3.82
+    - {{ compiler('c') }}          # [not win]
+    - {{ compiler('cxx') }}        # [not win]
+    - m2w64-gcc                    # [win]


I think there's a compiler function for m2w64 too. See here.

You need to rerender the recipe for this to have an effect.

Hm, I already tried this before: - {{ compiler('m2w64_cxx') }} # [win] but it didn't work (c639a13). Argh, no use checking the log, when does CI start throwing away logs??
Come to think of it, I don't think I re-rendered as I did not know that that's a prerequisite, I'll try this again...

bilderbuchi · 2021-08-25T14:59:46Z

It looks like configure should handle this but it does not.

Yeah, right? That's what I thought, too. That USE_WIN32API thing looks promising!

It's funny, mingw-64 actually has gdb available (in fact, that's what I'm using in the meantime, but not installed via the conda ecosystem), I just can't find out how the manage to build it... :-/

bilderbuchi · 2021-08-25T17:39:44Z

hm, this could be the build script...

Otherwise, the gdb configure script does not correctly choose ws2tcpip.h instead of netdb.h et al.

jaimergp · 2021-08-25T18:00:16Z

I think USE_WIN32API needs to be a preprocessor variable (CFLAGS containing -DUSE_WIN32API), not an env var, but it's been a while since I have done Windows stuff.

edit -- hah, ok, you pushed that commit as I posted this :D

bilderbuchi · 2021-08-25T18:00:51Z

I'm currently on it :D

bilderbuchi · 2021-08-25T18:27:27Z

After waiting for the sacred Interval, you turn another page in the mysterious book. Like the tomes around you in the great library, it is full of inscrutable and arcane symbols. As the sparkling dust settles, you glimpse yet another foe. The horror that arises from the page is...

checking whether strtoul is declared...   CXXLD  gdbreplay.exe
yes
gdbreplay.o:gdbreplay.cc:(.text+0xf7): undefined reference to `__imp_closesocket'
gdbreplay.o:gdbreplay.cc:(.text+0x224): undefined reference to `__imp_WSAStartup'
gdbreplay.o:gdbreplay.cc:(.text+0x276): undefined reference to `__imp_getaddrinfo'
gdbreplay.o:gdbreplay.cc:(.text+0x294): undefined reference to `gai_strerrorA'
gdbreplay.o:gdbreplay.cc:(.text+0x35d): undefined reference to `__imp_socket'
gdbreplay.o:gdbreplay.cc:(.text+0x3d3): undefined reference to `__imp_setsockopt'
gdbreplay.o:gdbreplay.cc:(.text+0x482): undefined reference to `__imp_bind'
gdbreplay.o:gdbreplay.cc:(.text+0x4db): undefined reference to `__imp_listen'
gdbreplay.o:gdbreplay.cc:(.text+0x511): undefined reference to `__imp_accept'
gdbreplay.o:gdbreplay.cc:(.text+0x56a): undefined reference to `__imp_setsockopt'
gdbreplay.o:gdbreplay.cc:(.text+0x5a6): undefined reference to `__imp_setsockopt'
gdbreplay.o:gdbreplay.cc:(.text+0x5e2): undefined reference to `__imp_getnameinfo'
gdbreplay.o:gdbreplay.cc:(.text+0x641): undefined reference to `__imp_closesocket'
gdbreplay.o:gdbreplay.cc:(.rdata$.refptr.in6addr_any[.refptr.in6addr_any]+0x0): undefined reference to `in6addr_any'
../gdbsupport/libgdbsupport.a(netstuff.o):netstuff.cc:(.text+0x57): undefined reference to `__imp_freeaddrinfo'
collect2.exe: error: ld returned 1 exit status

⚔️ 🧙

Roll a D20 for monster lore... 12... it seems to be a Linking Demon!

bilderbuchi · 2021-08-25T18:37:44Z

@conda-forge-admin, please rerender

…a-forge-pinning 2021.08.25.17.17.57

bilderbuchi · 2021-08-25T20:23:23Z

OK, LDFLAGS seems to end up withe reasonable values 'LDFLAGS= -lwsock32 -lws2_32 -Wl,--stack,12582912', but still ld fails to find these winsock(?) symbols. Dunno what's going on right now.

phil-blain · 2021-08-25T20:59:43Z

You could try doing make V=1, this will output the full commands to the log file instead of CXX <file>.o, CXXLD <exe>, etc ... Also adding -O (capital O) to the make invocation might help with the intermingled lines (in theory it should, not sure it will work with the CI layer log redirection).

bilderbuchi · 2021-08-29T21:16:51Z

Thanks for those tips! This is the relevant log section

2021-08-29T20:44:46.4356997Z g++  -std=gnu++17    -I. -I/d/bld/gdb_1630268986132/work/gdbserver -I/d/bld/gdb_1630268986132/work/gdbserver/../gdb/regformats -I/d/bld/gdb_1630268986132/work/gdbserver/.. -I/d/bld/gdb_1630268986132/work/gdbserver/../include -I/d/bld/gdb_1630268986132/work/gdbserver/../gdb -I/d/bld/gdb_1630268986132/work/gdbserver/../gnulib/import -I../gnulib/import -I/d/bld/gdb_1630268986132/work/gdbserver/.. -I.. -I./../intl -I/d/bld/gdb_1630268986132/_h_env/include -DUSE_WIN32API -pthread -Wall -Wpointer-arith -Wno-unused -Wunused-value -Wunused-variable -Wunused-function -Wno-switch -Wno-char-subscripts -Wempty-body -Wunused-but-set-parameter -Wunused-but-set-variable -Wno-sign-compare -Wno-error=maybe-uninitialized -Wsuggest-override -Wmissing-declarations -Wstrict-null-sentinel -Wformat -Wformat-nonliteral  -DGDBSERVER -lwsock32 -lws2_32 -Wl,--stack,12582912  \
2021-08-29T20:44:46.4358761Z 	-o gdbreplay.exe gdbreplay.o utils.o version.o  \
2021-08-29T20:44:46.4359436Z 	../gdbsupport/libgdbsupport.a ../gnulib/import/libgnu.a ../libiberty/libiberty.a ./../intl/libintl.a -liconv \
2021-08-29T20:44:46.4360053Z 	
2021-08-29T20:44:46.4360641Z gdbreplay.o:gdbreplay.cc:(.text+0xf7): undefined reference to `__imp_closesocket'
2021-08-29T20:44:46.4361495Z gdbreplay.o:gdbreplay.cc:(.text+0x224): undefined reference to `__imp_WSAStartup'
2021-08-29T20:44:46.4362202Z gdbreplay.o:gdbreplay.cc:(.text+0x276): undefined reference to `__imp_getaddrinfo'
2021-08-29T20:44:46.4363098Z gdbreplay.o:gdbreplay.cc:(.text+0x294): undefined reference to `gai_strerrorA'
2021-08-29T20:44:46.4363768Z gdbreplay.o:gdbreplay.cc:(.text+0x35d): undefined reference to `__imp_socket'
2021-08-29T20:44:46.4364492Z gdbreplay.o:gdbreplay.cc:(.text+0x3d3): undefined reference to `__imp_setsockopt'
2021-08-29T20:44:46.4365190Z gdbreplay.o:gdbreplay.cc:(.text+0x482): undefined reference to `__imp_bind'
2021-08-29T20:44:46.4365855Z gdbreplay.o:gdbreplay.cc:(.text+0x4db): undefined reference to `__imp_listen'
2021-08-29T20:44:46.4366680Z gdbreplay.o:gdbreplay.cc:(.text+0x511): undefined reference to `__imp_accept'
2021-08-29T20:44:46.4367391Z gdbreplay.o:gdbreplay.cc:(.text+0x56a): undefined reference to `__imp_setsockopt'
2021-08-29T20:44:46.4368068Z gdbreplay.o:gdbreplay.cc:(.text+0x5a6): undefined reference to `__imp_setsockopt'
2021-08-29T20:44:46.4368789Z gdbreplay.o:gdbreplay.cc:(.text+0x5e2): undefined reference to `__imp_getnameinfo'
2021-08-29T20:44:46.4369609Z gdbreplay.o:gdbreplay.cc:(.text+0x641): undefined reference to `__imp_closesocket'
2021-08-29T20:44:46.4370351Z gdbreplay.o:gdbreplay.cc:(.rdata$.refptr.in6addr_any[.refptr.in6addr_any]+0x0): undefined reference to `in6addr_any'
2021-08-29T20:44:46.4371166Z ../gdbsupport/libgdbsupport.a(netstuff.o):netstuff.cc:(.text+0x57): undefined reference to `__imp_freeaddrinfo'
2021-08-29T20:44:46.4371878Z collect2.exe: error: ld returned 1 exit status
2021-08-29T20:44:46.4372503Z Makefile:362: recipe for target 'gdbreplay.exe' failed
2021-08-29T20:44:46.4373101Z make[2]: *** [gdbreplay.exe] Error 1

gdbreplay.o gets compiled here (apparently without mishap)

2021-08-29T20:44:28.5697995Z g++  -std=gnu++17    -I. -I/d/bld/gdb_1630268986132/work/gdbserver -I/d/bld/gdb_1630268986132/work/gdbserver/../gdb/regformats -I/d/bld/gdb_1630268986132/work/gdbserver/.. -I/d/bld/gdb_1630268986132/work/gdbserver/../include -I/d/bld/gdb_1630268986132/work/gdbserver/../gdb -I/d/bld/gdb_1630268986132/work/gdbserver/../gnulib/import -I../gnulib/import -I/d/bld/gdb_1630268986132/work/gdbserver/.. -I.. -I./../intl -I/d/bld/gdb_1630268986132/_h_env/include -DUSE_WIN32API -pthread -Wall -Wpointer-arith -Wno-unused -Wunused-value -Wunused-variable -Wunused-function -Wno-switch -Wno-char-subscripts -Wempty-body -Wunused-but-set-parameter -Wunused-but-set-variable -Wno-sign-compare -Wno-error=maybe-uninitialized -Wsuggest-override -Wmissing-declarations -Wstrict-null-sentinel -Wformat -Wformat-nonliteral  -DGDBSERVER -c -o gdbreplay.o -MT gdbreplay.o -MMD -MP -MF ./.deps/gdbreplay.Tpo /d/bld/gdb_1630268986132/work/gdbserver/gdbreplay.cc

The missing references (WSAStartup etc) shown above are behind #ifdef USE_WIN32API guards in gdbreplay.c, so the -DUSE_WIN32API seems to work in order. It just cannot seem to find the library linked with -lwsock32 -lws2_32, I don't know why.

phil-blain · 2021-08-30T14:31:17Z

Copying the linking command with some newlines and commands added:

g++  -std=gnu++17    \
# Includes
-I. -I/d/bld/gdb_1630268986132/work/gdbserver \
-I/d/bld/gdb_1630268986132/work/gdbserver/../gdb/regformats \
-I/d/bld/gdb_1630268986132/work/gdbserver/.. \
-I/d/bld/gdb_1630268986132/work/gdbserver/../include \
-I/d/bld/gdb_1630268986132/work/gdbserver/../gdb \
-I/d/bld/gdb_1630268986132/work/gdbserver/../gnulib/import \
-I../gnulib/import -I/d/bld/gdb_1630268986132/work/gdbserver/.. \
-I.. -I./../intl -I/d/bld/gdb_1630268986132/_h_env/include \
# Preprocessor
-DUSE_WIN32API -pthread \
# Warnings
-Wall -Wpointer-arith -Wno-unused -Wunused-value -Wunused-variable -Wunused-function -Wno-switch -Wno-char-subscripts -Wempty-body -Wunused-but-set-parameter -Wunused-but-set-variable -Wno-sign-compare -Wno-error=maybe-uninitialized -Wsuggest-override -Wmissing-declarations -Wstrict-null-sentinel -Wformat -Wformat-nonliteral  \
# More preprocessor, your added LDFLAGS
-DGDBSERVER -lwsock32 -lws2_32 -Wl,--stack,12582912  \
# Object files
-o gdbreplay.exe gdbreplay.o utils.o version.o  \
# Static libs
../gdbsupport/libgdbsupport.a ../gnulib/import/libgnu.a ../libiberty/libiberty.a ./../intl/libintl.a \
# Shared libs
-liconv

According to this SO answer, the -lwsock32 -lws2_32 have to come after the object files. I think that's what you'd want to achieve, somehow.

Looking briefly at the code, I think this should all work correctly, without having to set USE_WIN32API manually.
Here is an excerpt from gdb/configure:

  WIN32APILIBS=
  case ${host} in
    *mingw32*)

$as_echo "#define USE_WIN32API 1" >>confdefs.h

      WIN32APILIBS="-lws2_32"
      ;;
  esac

So I think the problem is that host is not what this configure bit expects, in fact configure says it's: x86_64-pc-mingw64 (which I think is correct). So maybe a simple fix would be to patch this configure (and maybe others) to use *mingw* instead of *mingw32* ?

bilderbuchi · 2023-09-25T09:04:42Z

I'm closing this, this is hopelessly stale by now. Someone else, feel free to reopen/pick up where I left.

bilderbuchi requested review from gqmelo, marcelotrevisani and phil-blain as code owners June 14, 2021 15:23

bilderbuchi changed the title ~~Windows build~~ [WIP] Windows build Jun 14, 2021

bilderbuchi marked this pull request as draft June 14, 2021 18:26

bilderbuchi force-pushed the windows-build branch from 79aa8d5 to f360ad1 Compare June 14, 2021 19:47

bilderbuchi mentioned this pull request Jun 25, 2021

Update to GDB 10.2 #43

Merged

4 tasks

Enable Windows builds.

9a53d85

Start bld.bat using snippets from AnacondaRecipes [1] The script calls build.sh to reduce duplication. Remove packages unavailable on Windows. Add necessary m2 packages. [1] https://github.com/AnacondaRecipes/conda-feedstock/blob/master/recipe/bld.bat

bilderbuchi force-pushed the windows-build branch from 1616548 to 9a53d85 Compare July 1, 2021 19:14

conda-forge-linter and others added 2 commits July 1, 2021 19:16

MNT: Re-rendered with conda-build 3.21.4, conda-smithy 3.10.3, and co…

181521f

…nda-forge-pinning 2021.07.01.17.09.14

Add m2-ncurses on Windows.

52f8f5f

bilderbuchi added 4 commits July 5, 2021 09:18

Use m2-ncurses as a Windows run dependency

46cb64c

Avoid "unbound variable" error on Windows.

c7b9e2d

Echo build.sh lines for more debugging detail.

280061b

Use TMPDIR only on osx

da6b081

Avoids a warning on Windows build.

Don't use custom debugdir path on Windows.

e37c8c5

bilderbuchi force-pushed the windows-build branch from 8ccfac9 to e37c8c5 Compare August 21, 2021 10:54

jaimergp reviewed Aug 25, 2021

View reviewed changes

Try manually choosing WIN32API.

8ccd620

Otherwise, the gdb configure script does not correctly choose ws2tcpip.h instead of netdb.h et al.

bilderbuchi force-pushed the windows-build branch from a1bc208 to 8ccd620 Compare August 25, 2021 18:00

Let's see if the jinja compiler function works if we re-render.

770987b

MNT: Re-rendered with conda-build 3.21.4, conda-smithy 3.12, and cond…

642aaa1

…a-forge-pinning 2021.08.25.17.17.57

bilderbuchi force-pushed the windows-build branch from 479401f to 44d7bd6 Compare August 25, 2021 19:19

Link to the Winsock library.

0a02f55

bilderbuchi force-pushed the windows-build branch from 44d7bd6 to 0a02f55 Compare August 25, 2021 19:56

Tweak make args to maybe get better logs.

1727a04

bilderbuchi closed this Sep 25, 2023

[WIP] Windows build #42

[WIP] Windows build #42

Conversation

bilderbuchi commented Jun 14, 2021 • edited Loading

conda-forge-linter commented Jun 14, 2021

bilderbuchi commented Jun 14, 2021

conda-forge-linter commented Jun 15, 2021

conda-forge-linter commented Jun 15, 2021

bilderbuchi commented Jun 15, 2021 • edited Loading

bilderbuchi commented Jul 1, 2021

bilderbuchi commented Jul 1, 2021 • edited Loading

bilderbuchi commented Jul 1, 2021

klauer commented Jul 1, 2021

bilderbuchi commented Jul 2, 2021

bilderbuchi commented Jul 2, 2021 • edited Loading

github-actions bot commented Jul 2, 2021

github-actions bot commented Jul 2, 2021

github-actions bot commented Jul 2, 2021

phil-blain commented Jul 3, 2021

bilderbuchi commented Aug 3, 2021 • edited Loading

bilderbuchi commented Aug 3, 2021

bilderbuchi commented Aug 20, 2021 • edited Loading

bilderbuchi commented Aug 21, 2021 • edited Loading

bilderbuchi commented Aug 24, 2021

bilderbuchi commented Aug 24, 2021 • edited Loading

jaimergp commented Aug 25, 2021 • edited Loading

jaimergp Aug 25, 2021

Choose a reason for hiding this comment

jaimergp Aug 25, 2021

Choose a reason for hiding this comment

bilderbuchi Aug 25, 2021 • edited Loading

Choose a reason for hiding this comment

bilderbuchi commented Aug 25, 2021

bilderbuchi commented Aug 25, 2021

jaimergp commented Aug 25, 2021 • edited Loading

bilderbuchi commented Aug 25, 2021

bilderbuchi commented Aug 25, 2021 • edited Loading

bilderbuchi commented Aug 25, 2021

bilderbuchi commented Aug 25, 2021

phil-blain commented Aug 25, 2021 • edited Loading

bilderbuchi commented Aug 29, 2021 • edited Loading

phil-blain commented Aug 30, 2021

bilderbuchi commented Sep 25, 2023 • edited Loading

bilderbuchi commented Jun 14, 2021 •

edited

Loading

bilderbuchi commented Jun 15, 2021 •

edited

Loading

bilderbuchi commented Jul 1, 2021 •

edited

Loading

bilderbuchi commented Jul 2, 2021 •

edited

Loading

bilderbuchi commented Aug 3, 2021 •

edited

Loading

bilderbuchi commented Aug 20, 2021 •

edited

Loading

bilderbuchi commented Aug 21, 2021 •

edited

Loading

bilderbuchi commented Aug 24, 2021 •

edited

Loading

jaimergp commented Aug 25, 2021 •

edited

Loading

bilderbuchi Aug 25, 2021 •

edited

Loading

jaimergp commented Aug 25, 2021 •

edited

Loading

bilderbuchi commented Aug 25, 2021 •

edited

Loading

phil-blain commented Aug 25, 2021 •

edited

Loading

bilderbuchi commented Aug 29, 2021 •

edited

Loading

bilderbuchi commented Sep 25, 2023 •

edited

Loading