Linux, C++ and open source enthusiast.
- User Since
- Aug 17 2021, 5:38 PM (9 w, 2 d)
Fri, Sep 24
Also, don't we want to change the title and summary of the commit/revision? Because it does not reflect the real changes now in the repo
Thanks for the merge, I now understand more how this all works and what we want from these scripts.
Sep 9 2021
@MyDeveloperDay is it all good? I'm sorry for all the misunderstanding,
Sep 6 2021
only print info on generated plurals if there is a new plural
remove one print l17, tell me if we don't want the second print (now l17)
Sep 5 2021
Ok, so I removed git invocation and I tell the user what are their options, at least warnings are emitted for new plurals, and they will be shown in git status/diff and in revisions.
remove git invocation and user input, tell the user what the plurals are used for
Sep 4 2021
use correct python assignment from tuple, ask the user if they want to invoke git, use call instead of check_call to allow testing
Aug 31 2021
Aug 30 2021
add common plural rules, use python3 explicitly, reorder imports
generate plurals, for now supporting -y ending (-y to -ies/-ys), track generated plurals and show new ones to be checked
Ok, here is my proposal for plurals, I have some ideas, I think the safest/most complete would be to have a file tracking generated plurals and tell the user of the script to check newly generated plurals then add them to git, this would also allow reviewers to see new plurals generated. I'll put the updated revision of my idea soon.
Aug 27 2021
look at the html
For now, I handle plural manually, but it can be changed, I also kept Unsigned, what are your thoughts?
Use yaml type style for clang-format documentation (String, Integer, List of Strings, ...) instead of c++ types.
Ok, well, the reason I proposed this patch in the first place was that I have been working on a .clang-format schema (https://json-schema.org/) :) and I spotted the inconsistency. I checked, and clang-format reports an error if we give a negative value to an option expecting an unsigned. In the schema I am able to specify a minimum and I think it's appropriate to give the information to the user that it expect a positive/unsigned integer, what do you think?
Aug 26 2021
What about: unsigned? And enum or struct types such as BracketAlignmentStyle, ArrayInitializerAlignmentStyle, ...? Should we add something like: Enum BracketAlignmentStyle and Dictionnary BraceWrapping?
Thank you for your explanations, I understand now.