[src] changes to make kaldi can be installed on cygwin by qbetterk · Pull Request #2304 · kaldi-asr/kaldi

qbetterk · 2018-03-23T03:27:03Z

mostly the job is to add "defined(CYGWIN)" when needed and choose correct function as compiled on cygwin.

One thing worth noticing is that there is a line of code should be changed in "/usr/include/ctype.h". (still the problem of add "defined(CYGWIN)"). But this seems a problem of cygwin itself. Maybe user need to add it manually.

…erge

danpovey · 2018-03-23T18:12:42Z

src/base/kaldi-utils.cc

 #if defined(_MSC_VER) || defined(MINGW)
  ::Sleep(static_cast<int>(seconds * 1000.0));
+#elif defined(__CYGWIN__)
+  kaldi::Sleep(static_cast<int>(seconds * 1000.0));


this makes me think you haven't run the tests ("make test"). It would just recurse and crash when it exceeds the max stack depth.
Try doing "man usleep" or "man -k sleep" in cygwin, assuming man pages are installed, to see what sleep functions are available and what you have to include to find them.

jtrmal · 2018-03-28T18:26:09Z

src/util/kaldi-io.cc


+#ifdef __CYGWIN__
+static pclose(FILE* file) {
+  //FILE *file = (FILE*)_file;


I don't see the purpose of this function -- seems almost like a recursive call or somethign
y.

Sorry, this change(about 'popen' and 'pclose') is a mistake. The original code was right. I find that what I need to change is in the /usr/include/stdio.h rather than in kaldi.

qbetterk · 2018-03-30T20:00:21Z

@hainan-xv Can you review this?

jtrmal · 2018-03-30T20:32:49Z

src/base/get_version.sh

 # Overwrite ./version.h with the temporary file if they are different.
 if ! cmp -s $temp version.h; then
  cp $temp version.h
+  sed -i 's/\r//g' version.h


try to avoid sed -i (or sed altogether)

actually, try to figure out where the linefeed character comes from originally (does 'echo' add it?) and I'd like to know what problems it causes.

but there would be '\r' in the 'version.h' file since it is on windows. should I leave it to users to delete '\r' themselves?

kaldi itself does not have '\r'. this character is automatically generated when I download kaldi to a windows machine. '\r' in windows means 'end of line' like '\n' in Unix and the Unix environment(cygwin) cannot recognize it.

Actually, I think the logic of your change is wrong, your code will cause
the if ! cmd test always succeed, so you will always replace the file --
you should modify the $temp file instead

jtrmal · 2018-03-30T20:39:29Z

LGTM (besides the sed issue) BTW, there was (used to be) some compatibility feature to make the commands actiing like file (ie. those "command|command2|") specifiers work correctly within CYGWIN, make sure you didn't break those by adding/changing the ifdefs -- I think that relates to kaldi-io file. Also, did you run all tests successfully (`make test`)? y

…

On Fri, Mar 30, 2018 at 4:00 PM, Kun Qian ***@***.***> wrote: @hainan-xv <https://github.com/hainan-xv> Can you review this? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisX-NnVUuWY-CG_MTWjgCq8ztaT-H6ks5tjo7cgaJpZM4S4HvE> .

danpovey · 2018-03-30T20:41:43Z

Yes, testing should include running a setup like "yesno" on cygwin.

…

On Fri, Mar 30, 2018 at 4:39 PM, jtrmal ***@***.***> wrote: LGTM (besides the sed issue) BTW, there was (used to be) some compatibility feature to make the commands actiing like file (ie. those "command|command2|") specifiers work correctly within CYGWIN, make sure you didn't break those by adding/changing the ifdefs -- I think that relates to kaldi-io file. Also, did you run all tests successfully (`make test`)? y On Fri, Mar 30, 2018 at 4:00 PM, Kun Qian ***@***.***> wrote: > @hainan-xv <https://github.com/hainan-xv> Can you review this? > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#2304 (comment)>, or mute > the thread > <https://github.com/notifications/unsubscribe-auth/AKisX-NnVUuWY-CG_ MTWjgCq8ztaT-H6ks5tjo7cgaJpZM4S4HvE> > . > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADJVu_SH9AITfhUi_P0P6ETVvxnNGx6kks5tjpgFgaJpZM4S4HvE> .

jtrmal · 2018-03-30T20:55:21Z

Actually, I think the logic of your change is wrong, your code will cause the if ! cmd code always succeed, so you will always replace the file -- you should modify the $temp file instead y.

…

On Fri, Mar 30, 2018 at 4:42 PM, Kun Qian ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In src/base/get_version.sh <#2304 (comment)>: > @@ -89,6 +89,7 @@ fi # Overwrite ./version.h with the temporary file if they are different. if ! cmp -s $temp version.h; then cp $temp version.h + sed -i 's/\r//g' version.h but there would be '\r' in the 'version.h' file since it is on windows. should I leave it to users to delete '\r' themselves? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisX2bleM_OZCvX8p7i-ZijCYSqZehmks5tjpjNgaJpZM4S4HvE> .

jtrmal · 2018-03-30T20:56:20Z

does gcc/g++ mind the \r character there? y.

…

On Fri, Mar 30, 2018 at 4:54 PM, Jan Trmal ***@***.***> wrote: Actually, I think the logic of your change is wrong, your code will cause the if ! cmd code always succeed, so you will always replace the file -- you should modify the $temp file instead y. On Fri, Mar 30, 2018 at 4:42 PM, Kun Qian ***@***.***> wrote: > ***@***.**** commented on this pull request. > ------------------------------ > > In src/base/get_version.sh > <#2304 (comment)>: > > > @@ -89,6 +89,7 @@ fi > # Overwrite ./version.h with the temporary file if they are different. > if ! cmp -s $temp version.h; then > cp $temp version.h > + sed -i 's/\r//g' version.h > > but there would be '\r' in the 'version.h' file since it is on windows. > should I leave it to users to delete '\r' themselves? > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#2304 (comment)>, or mute > the thread > <https://github.com/notifications/unsubscribe-auth/AKisX2bleM_OZCvX8p7i-ZijCYSqZehmks5tjpjNgaJpZM4S4HvE> > . >

qbetterk · 2018-03-30T20:57:34Z

No, gcc/g++ seems ok with \r Sent from Mail for Windows 10 From: jtrmal Sent: Friday, March 30, 2018 4:56 PM To: kaldi-asr/kaldi Cc: Kun Qian; Author Subject: Re: [kaldi-asr/kaldi] [src] changes to make kaldi can be installed oncygwin (#2304) does gcc/g++ mind the \r character there? y.

On Fri, Mar 30, 2018 at 4:54 PM, Jan Trmal ***@***.***> wrote: Actually, I think the logic of your change is wrong, your code will cause the if ! cmd code always succeed, so you will always replace the file -- you should modify the $temp file instead y. On Fri, Mar 30, 2018 at 4:42 PM, Kun Qian ***@***.***> wrote: > ***@***.**** commented on this pull request. > ------------------------------ > > In src/base/get_version.sh > <#2304 (comment)>: > > > @@ -89,6 +89,7 @@ fi > # Overwrite ./version.h with the temporary file if they are different. > if ! cmp -s $temp version.h; then > cp $temp version.h > + sed -i 's/\r//g' version.h > > but there would be '\r' in the 'version.h' file since it is on windows. > should I leave it to users to delete '\r' themselves? > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#2304 (comment)>, or mute > the thread > <https://github.com/notifications/unsubscribe-auth/AKisX2bleM_OZCvX8p7i-ZijCYSqZehmks5tjpjNgaJpZM4S4HvE> > . >

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

jtrmal · 2018-03-30T20:59:49Z

OK, sorry, in that case I don't understand if the edit is needed at all? I might be missing somehting, tho. y.

…

On Fri, Mar 30, 2018 at 4:57 PM, Kun Qian ***@***.***> wrote: No, gcc/g++ seems ok with \r Sent from Mail for Windows 10 From: jtrmal Sent: Friday, March 30, 2018 4:56 PM To: kaldi-asr/kaldi Cc: Kun Qian; Author Subject: Re: [kaldi-asr/kaldi] [src] changes to make kaldi can be installed oncygwin (#2304) does gcc/g++ mind the \r character there? y. On Fri, Mar 30, 2018 at 4:54 PM, Jan Trmal ***@***.***> wrote: > Actually, I think the logic of your change is wrong, your code will cause > the if ! cmd code always succeed, so you will always replace the file -- > you should modify the $temp file instead > y. > > > On Fri, Mar 30, 2018 at 4:42 PM, Kun Qian ***@***.***> > wrote: > >> ***@***.**** commented on this pull request. >> ------------------------------ >> >> In src/base/get_version.sh >> <#2304 (comment)>: >> >> > @@ -89,6 +89,7 @@ fi >> # Overwrite ./version.h with the temporary file if they are different. >> if ! cmp -s $temp version.h; then >> cp $temp version.h >> + sed -i 's/\r//g' version.h >> >> but there would be '\r' in the 'version.h' file since it is on windows. >> should I leave it to users to delete '\r' themselves? >> >> — >> You are receiving this because you commented. >> Reply to this email directly, view it on GitHub >> <#2304 (comment)>, or mute >> the thread >> <https://github.com/notifications/unsubscribe- auth/AKisX2bleM_OZCvX8p7i-ZijCYSqZehmks5tjpjNgaJpZM4S4HvE> >> . >> > > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisX4j-lWfRftTZGUIvPdyvPlGRSimwks5tjpxPgaJpZM4S4HvE> .

qbetterk · 2018-03-30T21:00:01Z

Yes, I have run the ‘make test’ and there is no error. Kun From: jtrmal Sent: Friday, March 30, 2018 4:39 PM To: kaldi-asr/kaldi Cc: Kun Qian; Author Subject: Re: [kaldi-asr/kaldi] [src] changes to make kaldi can be installed oncygwin (#2304) LGTM (besides the sed issue) BTW, there was (used to be) some compatibility feature to make the commands actiing like file (ie. those "command|command2|") specifiers work correctly within CYGWIN, make sure you didn't break those by adding/changing the ifdefs -- I think that relates to kaldi-io file. Also, did you run all tests successfully (`make test`)? y

On Fri, Mar 30, 2018 at 4:00 PM, Kun Qian ***@***.***> wrote: @hainan-xv <https://github.com/hainan-xv> Can you review this? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisX-NnVUuWY-CG_MTWjgCq8ztaT-H6ks5tjo7cgaJpZM4S4HvE> .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

hainan-xv · 2018-04-02T19:17:08Z

src/fstext/lattice-weight-test.cc

+  } else if (tmp == 1) {
+  	return LatticeWeight( 1, 2);  // sometimes return special values..
+  } else if (tmp == 2) {
+  	return LatticeWeight( 2, 1);  // this tests more thoroughly certain properties...


you're not supposed to use \tab in the code; use 2 spaces instead.

jtrmal · 2018-04-02T19:23:41Z

Yes, please read the Google Style Guide for C++

qbetterk · 2018-04-02T19:59:04Z

Thanks! This guide seems useful! Best, Kun From: jtrmal Sent: Monday, April 2, 2018 3:25 PM To: kaldi-asr/kaldi Cc: Kun Qian; Author Subject: Re: [kaldi-asr/kaldi] [src] changes to make kaldi can be installed oncygwin (#2304) Yes, please read the Google Style Guide for C++ — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

danpovey

A couple of comments. please check with (e.g.) Hainan if you're unsure what I mean or how to do something.

danpovey · 2018-04-04T22:30:53Z

src/base/kaldi-utils.cc

 #if defined(_MSC_VER) || defined(MINGW)
  ::Sleep(static_cast<int>(seconds * 1000.0));
+#elif defined(__CYGWIN__)
+  sleep(static_cast<int>(seconds * 1.0));


the * 1.0 here is unnecessary.

danpovey · 2018-04-04T22:38:57Z

src/base/kaldi-utils.h

 #  define KALDI_MEMALIGN_FREE(x) _aligned_free(x)
+#elif defined(__CYGWIN__)
+#  define KALDI_MEMALIGN(align, size, pp_orig) \
+  (*(pp_orig) = aligned_alloc(align, size))


Actually I now see that we don't need this macro at all any more, as aligned_alloc is part of the C++11 standard. So please remove the macro and replace all instances of
foo = KALDI_MEMALIGN(x, y, &z)
with
foo = aligned_alloc(x, y);
and remove the temporary variables that previously bound to z, and replace KALDI_MEMALIGN_FREE with free (there are instances of this in the cudamatrix/ directory as well, that should also be replaced).

qbetterk · 2018-04-04T23:59:11Z

OK I would change it. By the way, the test of ‘yesno’ on both Cygwin and Linux are all finished with no error. Best, Kun

…

On Apr 4, 2018, at 18:52, Daniel Povey ***@***.***> wrote: @danpovey commented on this pull request. A couple of comments. please check with (e.g.) Hainan if you're unsure what I mean or how to do something. In src/base/kaldi-utils.cc: > @@ -45,6 +45,8 @@ std::string CharToString(const char &c) { void Sleep(float seconds) { #if defined(_MSC_VER) || defined(MINGW) ::Sleep(static_cast<int>(seconds * 1000.0)); +#elif defined(__CYGWIN__) + sleep(static_cast<int>(seconds * 1.0)); the * 1.0 here is unnecessary. In src/base/kaldi-utils.h: > # define KALDI_MEMALIGN(align, size, pp_orig) \ (*(pp_orig) = _aligned_malloc(size, align)) # define KALDI_MEMALIGN_FREE(x) _aligned_free(x) +#elif defined(__CYGWIN__) +# define KALDI_MEMALIGN(align, size, pp_orig) \ + (*(pp_orig) = aligned_alloc(align, size)) Actually I now see that we don't need this macro at all any more, as aligned_alloc is part of the C++11 standard. So please remove the macro and replace all instances of foo = KALDI_MEMALIGN(x, y, &z) with foo = aligned_alloc(x, y); and remove the temporary variables that previously bound to z, and replace KALDI_MEMALIGN_FREE with free (there are instances of this in the cudamatrix/ directory as well, that should also be replaced). — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

jtrmal · 2018-04-05T00:15:38Z

and do they have comparable scores? y.

…

On Wed, Apr 4, 2018 at 7:59 PM, Kun Qian ***@***.***> wrote: OK I would change it. By the way, the test of ‘yesno’ on both Cygwin and Linux are all finished with no error. Best, Kun > On Apr 4, 2018, at 18:52, Daniel Povey ***@***.***> wrote: > > @danpovey commented on this pull request. > > A couple of comments. please check with (e.g.) Hainan if you're unsure what I mean or how to do something. > > In src/base/kaldi-utils.cc: > > > @@ -45,6 +45,8 @@ std::string CharToString(const char &c) { > void Sleep(float seconds) { > #if defined(_MSC_VER) || defined(MINGW) > ::Sleep(static_cast<int>(seconds * 1000.0)); > +#elif defined(__CYGWIN__) > + sleep(static_cast<int>(seconds * 1.0)); > the * 1.0 here is unnecessary. > > In src/base/kaldi-utils.h: > > > # define KALDI_MEMALIGN(align, size, pp_orig) \ > (*(pp_orig) = _aligned_malloc(size, align)) > # define KALDI_MEMALIGN_FREE(x) _aligned_free(x) > +#elif defined(__CYGWIN__) > +# define KALDI_MEMALIGN(align, size, pp_orig) \ > + (*(pp_orig) = aligned_alloc(align, size)) > Actually I now see that we don't need this macro at all any more, as aligned_alloc is part of the C++11 standard. So please remove the macro and replace all instances of > foo = KALDI_MEMALIGN(x, y, &z) > with > foo = aligned_alloc(x, y); > and remove the temporary variables that previously bound to z, and replace KALDI_MEMALIGN_FREE with free (there are instances of this in the cudamatrix/ directory as well, that should also be replaced). > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub, or mute the thread. > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisXzxf77lNDaZvK693IsmrF6JzTenrks5tlV5WgaJpZM4S4HvE> .

qbetterk · 2018-04-05T00:47:11Z

I remember it is 0% WER. From: jtrmal Sent: Wednesday, April 4, 2018 8:16 PM To: kaldi-asr/kaldi Cc: Kun Qian; Author Subject: Re: [kaldi-asr/kaldi] [src] changes to make kaldi can be installed oncygwin (#2304) and do they have comparable scores? y.

On Wed, Apr 4, 2018 at 7:59 PM, Kun Qian ***@***.***> wrote: OK I would change it. By the way, the test of ‘yesno’ on both Cygwin and Linux are all finished with no error. Best, Kun > On Apr 4, 2018, at 18:52, Daniel Povey ***@***.***> wrote: > > @danpovey commented on this pull request. > > A couple of comments. please check with (e.g.) Hainan if you're unsure what I mean or how to do something. > > In src/base/kaldi-utils.cc: > > > @@ -45,6 +45,8 @@ std::string CharToString(const char &c) { > void Sleep(float seconds) { > #if defined(_MSC_VER) || defined(MINGW) > ::Sleep(static_cast<int>(seconds * 1000.0)); > +#elif defined(__CYGWIN__) > + sleep(static_cast<int>(seconds * 1.0)); > the * 1.0 here is unnecessary. > > In src/base/kaldi-utils.h: > > > # define KALDI_MEMALIGN(align, size, pp_orig) \ > (*(pp_orig) = _aligned_malloc(size, align)) > # define KALDI_MEMALIGN_FREE(x) _aligned_free(x) > +#elif defined(__CYGWIN__) > +# define KALDI_MEMALIGN(align, size, pp_orig) \ > + (*(pp_orig) = aligned_alloc(align, size)) > Actually I now see that we don't need this macro at all any more, as aligned_alloc is part of the C++11 standard. So please remove the macro and replace all instances of > foo = KALDI_MEMALIGN(x, y, &z) > with > foo = aligned_alloc(x, y); > and remove the temporary variables that previously bound to z, and replace KALDI_MEMALIGN_FREE with free (there are instances of this in the cudamatrix/ directory as well, that should also be replaced). > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub, or mute the thread. > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2304 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKisXzxf77lNDaZvK693IsmrF6JzTenrks5tlV5WgaJpZM4S4HvE> .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

…fixes kaldi-asr#2347 This reverts commit 72c32fd.

…fixes kaldi-asr#2347

…2347 (#2348)

…ros (kaldi-asr#2304)

…fixes kaldi-asr#2347 (kaldi-asr#2348)

…ros (kaldi-asr#2304)

…fixes kaldi-asr#2347 (kaldi-asr#2348)

Qian Kun added 6 commits March 13, 2018 22:01

fix generat_plot.py file to suit oderiv-rms value

c656652

fix plot generate files to suit oderiv-rms output

c662c5b

Merge branch 'master' of https://github.com/kaldi-asr/kaldi into to-m…

19db308

…erge

update chage for cygwin installation

4520ce4

update changes for kaldi installation on cygwin

6804629

last two changes for installing kaldi on cygwin

8c73204

danpovey reviewed Mar 23, 2018

View reviewed changes

jtrmal reviewed Mar 28, 2018

View reviewed changes

Qian Kun added 2 commits March 30, 2018 14:49

fix for the recursive problem and memory-allocating problem

1d41c08

missing changes

fae1dd9

jtrmal reviewed Mar 30, 2018

View reviewed changes

Qian Kun added 3 commits March 30, 2018 17:11

remove sed -i

164fc8d

precision problem

332d944

precision problem

1dcd0ff

hainan-xv reviewed Apr 2, 2018

View reviewed changes

change \tab to 2 spaces

fccd864

danpovey reviewed Apr 4, 2018

View reviewed changes

replace KALDI_MEMALIGN with aligned_alloc and related commond

72c32fd

danpovey merged commit 1a1e265 into kaldi-asr:master Apr 5, 2018

danpovey added a commit to danpovey/kaldi that referenced this pull request Apr 10, 2018

[src] Revert KALDI_MEMALIGN->aligned_alloc change from kaldi-asr#2304, …

74126b7

…fixes kaldi-asr#2347 This reverts commit 72c32fd.

danpovey added a commit to danpovey/kaldi that referenced this pull request Apr 10, 2018

[src] Revert KALDI_MEMALIGN->aligned_alloc change from kaldi-asr#2304, …

27e7c65

…fixes kaldi-asr#2347

danpovey added a commit that referenced this pull request Apr 10, 2018

[src] Revert KALDI_MEMALIGN->aligned_alloc change from #2304, fixes #…

e10c478

…2347 (#2348)

LvHang pushed a commit to LvHang/kaldi that referenced this pull request Apr 14, 2018

[src,build] Fix cygwin compatibility; remove unnecessary memalign mac…

c0c9561

…ros (kaldi-asr#2304)

LvHang pushed a commit to LvHang/kaldi that referenced this pull request Apr 14, 2018

[src] Revert KALDI_MEMALIGN->aligned_alloc change from kaldi-asr#2304, …

a9f5c8c

…fixes kaldi-asr#2347 (kaldi-asr#2348)

Skaiste pushed a commit to Skaiste/idlak that referenced this pull request Sep 26, 2018

[src,build] Fix cygwin compatibility; remove unnecessary memalign mac…

b0c493c

…ros (kaldi-asr#2304)

Skaiste pushed a commit to Skaiste/idlak that referenced this pull request Sep 26, 2018

[src] Revert KALDI_MEMALIGN->aligned_alloc change from kaldi-asr#2304, …

d95e00c

…fixes kaldi-asr#2347 (kaldi-asr#2348)

Conversation

qbetterk commented Mar 23, 2018

Uh oh!

danpovey Mar 23, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

qbetterk commented Mar 30, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jtrmal commented Mar 30, 2018 via email

Uh oh!

danpovey commented Mar 30, 2018 via email

Uh oh!

jtrmal commented Mar 30, 2018 via email

Uh oh!

jtrmal commented Mar 30, 2018 via email

Uh oh!

qbetterk commented Mar 30, 2018 via email

Uh oh!

jtrmal commented Mar 30, 2018 via email

Uh oh!

qbetterk commented Mar 30, 2018 via email

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jtrmal commented Apr 2, 2018

Uh oh!

qbetterk commented Apr 2, 2018 via email

Uh oh!

danpovey left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

qbetterk commented Apr 4, 2018 via email

Uh oh!

jtrmal commented Apr 5, 2018 via email

Uh oh!

qbetterk commented Apr 5, 2018 via email

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

danpovey Mar 23, 2018 •

edited

Loading