I've attached a pretty simple patch that fixes an item in
gnulib-tool.py.TODO. It fixes a lot of the remaining lines in the diff
between gnulib-tool and gnulib-tool.py which is nice.
Specifically, it fixes the CFLAGS value used for building the Gnulib
source files. Then it adds some extra rules to
Hello,
I have discovered the cause of the --add-import issue. It is simpler
than I originally thought. It seems that a code path is taken where
the value of 'sourcebase' in the GLConfig is not set. Then in
GLImport.rewrite_new_files this line is taken:
path = constants.substart('lib/', '%s/'
On 3/11/24 2:46 AM, Collin Funk wrote:
> So I think it is safe to say the problem occurs somewhere around
> there... All of those modules are Coreutils --local-dir modules. Then
> dynamic typing things happen.
I just started looking into this issue a bit more. With a clean
Coreutils directory, I g
Hi Bruno,
On 3/11/24 4:32 AM, Bruno Haible wrote:
> Another GNU package in the same realm is GNU Bison. If you like regular
> expression and automata theory, you will like GNU Bison as well.
Thanks for the suggestion. I've never gotten around to using YACC or
GNU Bison, surprisingly. I'll find so
On 3/11/24 5:37 AM, Bruno Haible wrote:
> There is a pitfall with the list.join('\n') idiom: it's mainly designed for
> non-empty lists. I therefore add the patch below.
Ah, I see what you mean. I didn't even think of this case...
Good catch. Thanks.
Collin
On 3/11/24 4:57 AM, Bruno Haible wrote:
> This looks good to me. 'sed' accepts POSIX BREs, where '(' and '[' stand for
> the literal '(' and '[' characters unless escaped.
Thanks for double checking for me.
> Also, I applied the attached follow-up, in order to make the logic of the code
> easier
Hi Collin,
> > - A function _eliminate_NMD_from_line that takes a single line as argument
> > and return a line or None.
> > - A function _eliminate_NMD that invokes _eliminate_NMD_from_line on each
> > line of the snippet, and combines the results?
>
> I tried to implement it as you
Hi Collin,
> On 3/10/24 5:13 AM, Bruno Haible wrote:
> > automake_options = {x for y in configure_ac_automake_options for x in
> > y.split()}
>
> I've attached the fixed patch.
Thanks, applied.
> I would like to double check the regular expression I used:
>
> pattern = re.compile(r'^.*A
Hi Collin,
> The POSIX BRE vs. ERE along with backreferences, assertions, lazy vs.
> greedy matching allowed by some languages adds multiple layers of
> complexity which starts to hurt my brain. :)
Yes, the many variants of regular expressions are complicated...
> The DFA module also looks prett
Hi Simon,
On 3/11/24 1:06 AM, Simon Josefsson wrote:
> Trying gnulib-tool.py on OATH Toolkit (which use a somewhat unorthodox
> gnulib usage style by adding code into git) results in error below. I
> thought it was a missing mkdir at some point, but I couldn't find a
> solution... ideas?
I was g
Hi Bruno,
I have attached two updated patches for 0003 and 0004.
On 3/10/24 6:18 AM, Bruno Haible wrote:
> In the function _eliminate_NMD, the list accumulation expressions are a bit
> complex.
Sorry, fixed. I am a big fan of Python's list comprehensions [1]. :)
I will concede that they can get
I like the plan to replace gnulib-tool with a faster implementation, and
a two-year migration phase sounds reasonable to see if it will work in
practice.
Trying gnulib-tool.py on OATH Toolkit (which use a somewhat unorthodox
gnulib usage style by adding code into git) results in error below. I
th
12 matches
Mail list logo
