Re: Vitalije: PR 2720 suggests improving Leo's python importer

On Wednesday, July 27, 2022 at 1:30:37 PM UTC-5 Edward K. Ream wrote: > On Wed, Jul 27, 2022 at 10:36 AM tbp1...@gmail.com > wrote > >> I think it is unwise to depend on some external program to detect Leo-specific errors. > git init will create the local repo. No need to push it anywhere els

Re: Vitalije: PR 2720 suggests improving Leo's python importer

On Wed, Jul 27, 2022 at 10:36 AM tbp1...@gmail.com wrote: > Not all imports come from a git repo. I would say that the majority of > mine (historically) have come from previously existing files on my system > that will never get into a repo. I think it is unwise to depend on some > external pr

Re: Vitalije: PR 2720 suggests improving Leo's python importer

Not all imports come from a git repo. I would say that the majority of mine (historically) have come from previously existing files on my system that will never get into a repo. I think it is unwise to depend on some external program to detect Leo-specific errors. On Wednesday, July 27, 2022

Re: Vitalije: PR 2720 suggests improving Leo's python importer

On Monday, July 25, 2022 at 7:28:17 AM UTC-5 Edward K. Ream wrote: P.S. I may alter Vitalije's importer slightly. At present, Vitalije's > importer carefully adds Leo's escape string for underindented lines. Imo, > it may be better *not *to add those escapes, even if Leo does not import > some

Re: Vitalije: PR 2720 suggests improving Leo's python importer

> It may be possible to simplify most of Leo's present importers using the ideas in Vitalije's python importer. Study and experimentation show: 1. It is not possible to create tokenizers from the existing scan tables. Creating a new tokenizer for each language is out of the question, so any n

Re: Vitalije: PR 2720 suggests improving Leo's python importer

>> Good thing that the importers get a 'redesign' and/or get simplified, as I do intend to transliterate them for LeoJS. > I plan only to retrofit the python importer. All the others should stay as they are. On second thought, I may try to simplify all the importers. PR #2729

Re: Vitalije: PR 2720 suggests improving Leo's python importer

> Good thing that the importers get a 'redesign' and/or get simplified, as I do intend to transliterate them for LeoJS. I plan only to retrofit the python importer. All the others should stay as they are. The base Importer class supports tables that are, in effect, language-specific tokenizers. I

Re: Vitalije: PR 2720 suggests improving Leo's python importer

Not sure i understand any of this other than superficially, as I have not even taken a look at the importer code ever :/ But, I just wanted to chime in to say: Good thing that the importers get a 'redesign' and/or get simplified, as I do intend to transliterate them for LeoJS. I still ha

Re: Vitalije: PR 2720 suggests improving Leo's python importer

On Sun, Jul 17, 2022 at 9:04 AM Edward K. Ream wrote: > PR #2720 is ready > for review. > Vitalije has signed off on the PR. I'll merge it soon. I plan to use Vitalije's code as the basis for yet another (experimental) importer, based on the

Vitalije: PR 2720 suggests improving Leo's python importer

PR #2720 is ready for review. Vitalije, I would appreciate your comments. I have spent several pleasant days reviewing every line of Vitalije's clever importer code. I believe I now understand it completely. The PR lists the changes. Most ar