flex package POSIX violation

flex package POSIX violation

[Steven Penny]

The POSIX standard, at least since 1997 [1], requires the presence of a "lex"
utility.

Most systems (including Cygwin) provide a Lex implementation via the "flex"
package. However with other OS, a "lex -> flex" symlink is provided, whereas the
Cygwin package does not.

Please resolve this, thanks.

[1] http://pubs.opengroup.org/onlinepubs/7908799/xcu/lex.html


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Marco Atzeri]

On 31/12/2017 22:13, Steven Penny wrote:
> The POSIX standard, at least since 1997 [1], requires the presence of a 
> "lex"
> utility.
> 
> Most systems (including Cygwin) provide a Lex implementation via the "flex"
> package. However with other OS, a "lex -> flex" symlink is provided, 
> whereas the
> Cygwin package does not.
> 
> Please resolve this, thanks.
> 
> [1] http://pubs.opengroup.org/onlinepubs/7908799/xcu/lex.html
> 
> 

Flex is NOT 100% Posix compliant.
See Info on
   20 Incompatibilities with Lex and Posix
as upstream does not provide the link "lex -> flex"
I am reluctant to do differently.

Regards
Marco

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Yaakov Selkowitz]

[-- Attachment #1.1: Type: text/plain, Size: 887 bytes --]

On 2017-12-31 15:44, Marco Atzeri wrote:
> On 31/12/2017 22:13, Steven Penny wrote:
>> The POSIX standard, at least since 1997 [1], requires the presence of
>> a "lex" utility.
>>
>> Most systems (including Cygwin) provide a Lex implementation via the
>> "flex"
>> package. However with other OS, a "lex -> flex" symlink is provided,
>> whereas the Cygwin package does not.
> 
> Flex is NOT 100% Posix compliant.
> See Info on
>   20 Incompatibilities with Lex and Posix
> as upstream does not provide the link "lex -> flex"
> I am reluctant to do differently.

This is generally handled downstream, e.g.:

http://pkgs.fedoraproject.org/cgit/rpms/flex.git/tree/flex.spec#n90
https://packages.debian.org/sid/amd64/flex/filelist
https://gitweb.gentoo.org/repo/gentoo.git/tree/sys-devel/flex/flex-2.6.4-r1.ebuild#n84

We should also do accordingly.

-- 
Yaakov


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: flex package POSIX violation

[Stephen John Smoogen]

On 31 December 2017 at 16:13, Steven Penny <svnpenn@gmail.com> wrote:
> The POSIX standard, at least since 1997 [1], requires the presence of a
> "lex"
> utility.
>
> Most systems (including Cygwin) provide a Lex implementation via the "flex"
> package. However with other OS, a "lex -> flex" symlink is provided, whereas
> the
> Cygwin package does not.
>
> Please resolve this, thanks.
>

How POSIX compliant is Cygwin supposed to be? I don't think many of
the tools are 100% POSIX compliant but are good enough so does making
the symlink between flex and lex make it more compliant or less so
because the tool doesn't meet 100% compliance?





-- 
Stephen J Smoogen.

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Gerrit Haase]

2017-12-31 22:13 GMT+01:00 Steven Penny writes:
> The POSIX standard, at least since 1997 [1], requires the presence of a
> "lex"
> utility.
>
> Most systems (including Cygwin) provide a Lex implementation via the "flex"
> package. However with other OS, a "lex -> flex" symlink is provided, whereas
> the
> Cygwin package does not.
>
> Please resolve this, thanks.
>
> [1] http://pubs.opengroup.org/onlinepubs/7908799/xcu/lex.html


Maybe to 'request' something is not the best approach to resolve your issue?

A better way to handle issues like this would be:
you provide a patch to handle the issue and the maintainer of the
package may then decide to incorporate your patch.


Thanks.

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Steven Penny]

On Fri, 5 Jan 2018 15:37:11, Gerrit Haase wrote:
> Maybe to 'request' something is not the best approach to resolve your issue?

yeah, sometimes it is. my original request was Dec 31 [1]. by Jan 1 [2], updated
package had been published.

> A better way to handle issues like this would be:
> you provide a patch to handle the issue and the maintainer of the
> package may then decide to incorporate your patch.

maybe you should read full threads before injecting your opinion. this advice
might hold true for larger changes - but we are literally talking about 1
symlink.

And even if that werent the case, this is open source. *I* decide my own level
of involvement, not you. if i want to make a half ass request with no effort,
that is my decision and it might take years or never for it to be implemented,
or it might be the next day as we see here. please look here [3] to see that i
am well versed on how to make a patch - cheers

[1] http://cygwin.com/ml/cygwin/2017-12/msg00296.html
[2] http://cygwin.com/ml/cygwin/2018-01/msg00004.html
[3] http://github.com/svnpenn


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Stephen John Smoogen]

On 31 December 2017 at 19:26, Steven Penny <svnpenn@gmail.com> wrote:
> On Sun, 31 Dec 2017 18:43:00, Stephen John Smoogen wrote:
>>
>> Which is why I wanted to see where this was going. Are these fixes
>> just looking for low hanging fruit to be POSIX compliant, or are these
>> needing larger amounts of resources to be 'compliant'? If the
>> flex->lex link fails some sort of POSIX test, are people going to need
>> Cygwin porters to fix those? Also is there an easy line for "this is
>> compliant enough?"
>
>
> Here is a simple demonstration of the problem:
>
>    $ cat xr.l
>    %option main
>    %%
>    ya printf("zu");
>    %%
>
>    $ make xr
>    lex  -t xr.l > xr.c
>    /bin/sh: lex: command not found
>    make: *** [<builtin>: xr.c] Error 127
>    rm xr.c
>
> now of course you can work around this by "make LEX=flex xr" or similar, but
> no
> major Linux distro makes you do this, as they already include "lex"
> vis-a-vis
> the symlink to flex.
>

OK that makes it a clearer and tangible problem to me. Thank you for
putting up with my questions.

> [1] http://gnu.org/software/make/manual/html_node/Implicit-Variables
>
>
>
> --
> Problem reports:       http://cygwin.com/problems.html
> FAQ:                   http://cygwin.com/faq/
> Documentation:         http://cygwin.com/docs.html
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
>



-- 
Stephen J Smoogen.

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: flex package POSIX violation

[Steven Penny]

On Sun, 31 Dec 2017 18:43:00, Stephen John Smoogen wrote:
> Which is why I wanted to see where this was going. Are these fixes
> just looking for low hanging fruit to be POSIX compliant, or are these
> needing larger amounts of resources to be 'compliant'? If the
> flex->lex link fails some sort of POSIX test, are people going to need
> Cygwin porters to fix those? Also is there an easy line for "this is
> compliant enough?"

Here is a simple demonstration of the problem:

    $ cat xr.l
    %option main
    %%
    ya printf("zu");
    %%

    $ make xr
    lex  -t xr.l > xr.c
    /bin/sh: lex: command not found
    make: *** [<builtin>: xr.c] Error 127
    rm xr.c

now of course you can work around this by "make LEX=flex xr" or similar, but no
major Linux distro makes you do this, as they already include "lex" vis-a-vis
the symlink to flex.

[1] http://gnu.org/software/make/manual/html_node/Implicit-Variables


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple