Documentation of FoxBin2Prg - A Binary to Text converter for MS Visual Foxpro 9
Using FoxPin2Prg with git.
See Use with SCM tools
As we all know, there is a whole bunch of tools that we might use to colaborate and to version our work.
Since I use git I try to point out what I do to do as fast and comfortable work.
This document gives no workflow, in special not a workflow of git. There is no wrong or right.
If you are not fimilar with git, please read through this first, there are minor settings that might help a bit on later work.
Please note, if you come from a SCM system, git feels very odd. It's a complete different idea of thinking. Do yourself the favour, on git-scm you find pro git. Read and work at least through the first three chapters. No matter, if you are using git on bash or GUI
For almost anything you don't need a server, even for colaborating a storage is enough. I use a server - a self hosted copy of gitlab - but this basically for access restrictions and remote work.
Do not hesitate to try anything. Always remember, git is just a couple of files inside your project folder. CTRL+C, CTRL+V on Explorer might save the day. ;)
A copy of FoxBin2Prg
It might be a good idea to have a copy of git.
If you download a setup or clone the code and compile yourself is up to you.
I prefer the setup - it's a nice package with all bells and whistles one might need on command line.
See git settings for stuff to set during setup.
I prefer it on the bash. Some of the reasons
- It's complex. If a GUI does something wrong, you need to go to the bash anyway.
- Note, every git-GUI is simple a tool that needs git to run original commands.
- It's simple. For day to day work, it's some easy to learn commands.
- The little visual need I have, and yes, some commands too, run on gitk, what ships with git
- bash, because all help out of the web comes in 'nix style, rethink to the windows world is to much work.
- working on the command line gives ideas of doing the stuff using the Fox, and we are programmmers, are we?
- since we have those extra step to create Text and Binary around the git repo, any GUI is awkward.
The use of git on the command line is not a taste all like.
So if you need a GUI tool, it's your choice. git-scm.org list a lot, and there might be more out.
Or use the Fox and build your own - it's nothing then some commands and handling text files.
To integrate git into the Fox VFPX lists some tools.
I'm somehow into the Bin 2 Text extension, but then I know this best.
A nice graphical compare / mergetool is nice too have - check if and how to integrate first, and also if the code mokey creating the interface understands git. Some try to use it like SCM, what ends up odd.
The first one must understand here, git is set up to deal with text based files.
It's a Linux tool developed to deal with the Kernel code.
Highly sophisticated, neat, fast.
So storing binaries into it is something one should avoid as far as possible.
There is stuff where on might not circumvent it - like pictures.
And there is the oddness of VFP sources. Storing those to git is as wrong as it could be.
- Bloat of the repo
- No trace of changes
- Not mergeable
And if there is one what is realy cool on skipping locking servers, it's colaboarting. Even on VCX's.
With thanks to Fernando FoxBin2Prg comes into play.
One might be reluctant to dropping the binaries - but I do for years and it works like a charm.
Do not play around with text-to-look-up-changes ideas storing both Text and Binary.
Just merge, compare like a normal text based language.
The only thing one must learn (and remember, even after odd merges) is to run FoxBin2prg wrapping any git operation.
git would even allow to hook FoxBin2Prg to run after some git operations, but this might be a killer.
I try to show what to set to keep it running as Text - only storage.
I know there is a lot how to store a class. Microseconds here and there.
Folks, hardware is cheap. There is no sense in counting ticks and bits any more. I'm grown up that way, but.
Dealing with the FoxBin2Prg.prg is the limit.
If I imagine to have one of my larger vcx's like this in a merge view - no.
Just split up the vcx's into single classes. It's like a charm on merge and diff.
Also on blame.
I you go for the lib.baseclass.classname.vc2 or lib.classname.vc2 style is a matter of taste.
I can see baseclass on classname, so there is not much value, other might think different.
But split into classes makes life easy.
Side effect is that SCX will be split into 3 files too. I have no idea if there is some use of it - I just do classes. If you like it separated from VCX - let me know.
Splitting DBC is possible too. My approach does not touch the DBC structure (it's meta data only) in the source over decades, so there is little use.
So I've added an option to inhibit splitting DBC while splitting VCX/SCX.
Storing as Text is a long term goal - the fork is about some of the problems.
Storing as Text is a long term goal - the fork is about some of the problems. Since nobody brought up the .NULL. problem, it looks like it's not commonly used. Depends on the data used. (And if there is any sense in merging.)
All the other table-based-sources of VFP work seamless as Text. No fuzz.
Assuming you do a fresh install of git, you will be asked questions over questions. The most you left unchanged.
The one I recommend is the dealing with line endings. You know all the odd uses of LF and CR.
git for windows offers to alter this while dealing with your files. I recommend not. Just "Check in as is, check out as is".
This will keep your files, and any modern editor or WEB UI will not care anyway.
If you missed on install - just install again.
For some odd reasons, the people on git-scm are going the way of evil. If you start the Setup of git, all depends who you are. If you not check Run as Adminstrator, it will install into your user folder. Even if installed in Programs folder earlier!
It's your taste to install 64 or 32 bit versions. Calling 64bit from the fox is odd, to say the least. If you don't have no other needs, use 32bit.
The .gitignore file controls which files go not into the git repository by default. It goes by folder with inheritance. If you undstand the inheritance of .gitignore, you grok FoXBin2Prg.cfg and vice versa. Very close.
But git would not be git if one could not teach .gitignore to define the files included. Again, see pro-git.
Creating the file might be a bit odd on MS Windows (Windows dislikes just extension). Created once, most editors will not complain.
In case, just open the bash in your projects base folder (Explorer, context menu) and enter touch .gitignore.
An example for Text only use.
#.gitignore
#exclude general
*.*
#include general
#dbx data
!*.dbf
!*.cdx
!*.fpt
!*.dbc
!*.dc[tx]
#grafics
!*.bmp
!*.msk
!*.ico
!*.cdr
!*.cur
#programms
!*.prg
!*.fpw
#header
!*.h
#libs (by FoxBin2prg)
#!*.vc[xt]
#!*.pj[xt]
#!*.fr[xt]
#!*.mn[xt]
#!*.sc[xt]
#!*.lb[xt]
#FoxBin2prg
!*.vc2
!*.pj2
!*.fr2
!*.mn2
!*.sc2
!*.lb2
#!*.d[bc]2
#include special
#default git
!*.gitignore
!/desktop.ini
#desktop.ini. usless? no! it keeeps the folder icon!
#diverse
!*.reg
!/FoxBin2prg.cfg
!README.md
#exclude special again
foxuser.*
_command.prg
As you see, this assumes that tables and databases are stored as binary.
I think I will swap to use them as Text too, since version 1.20.1 of FoxBin2Prg eleminates some problems with this idea.
Also this approach ignores any bak / log / tmp / err files by default.
The config of FoxBin2Prg using FoxBin2Prg.cfg files is relative simple. As you might see above, I carry the file within the repo.
The good on this is, if I change my mind and alter settings in it, FoxBin2PRG will find the setting that fit to the data on each commit.
If I have to checkout old stuff.
*################################################################################################################
*FOXBIN2PRG.CFG configuration options: (If no values given, these are the DEFAULTS)
*extension: tx2=newext && Specify extensions to use. Default FoxBin2Prg extensions ends in '2' (see at the bottom)
*ShowProgressbar: 1 && 0=Don't show, 1=Allways show, 2= Show only for multi-file processing
*DontShowErrors: 0 && Show message errors by default
*NoTimestamps: 1 && Clear timestamps by default for minimize differences
*Debug: 0 && Don't Activate individual <file>.Log by default
*BodyDevInfo: 0 && [0=Don't keep DevInfo for body pjx records], 1=Keep DevInfo
*ExtraBackupLevels: 1 && By default 1 BAK is created. With this you can make more .N.BAK, or none
*ClearUniqueID: 1 && 0=Keep UniqueID in text files, 1=Clear Unique ID. Useful for Diff and Merge
*ClearDBFLastUpdate: 1 && 0=Keep DBF LastUpdate, 1=Clear DBF LastUpdate. Useful for Diff.
*OptimizeByFilestamp: 0 && 1=Optimize file regeneration depending on file timestamp. Dangerous while working with branches!
*RemoveNullCharsFromCode: 1 && 1=Drop NULL chars from source code
*RemoveZOrderSetFromProps: 0 && 0=Do not remove ZOrderSet property from object, 1=Remove ZOrderSet property from object
*Language: (auto) && Language of shown messages and LOGs. EN=English, FR=French, ES=Español, DE=German, Not defined = AUTOMATIC [DEFAULT]
*ExcludeDBFAutoincNextval: 0 && [0=Do not exclude this value from db2], 1=Exclude this value from db2
*PRG_Compat_Level: 0 && [0=Legacy], 1=Use HELPSTRING as Class Procedure comment
*-- Convertion options:
*PJX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*VCX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*SCX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*FRX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*LBX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*MNX_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*FKY_Conversion_Support: 1 && 0=No support, 1=Generate TXT only (Diff)
*MEM_Conversion_Support: 1 && 0=No support, 1=Generate TXT only (Diff)
*DBC_Conversion_Support: 2 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
*DBF_Conversion_Support: 1 && 0=No support, 1=Generate Header TXT only (Diff), 2=Generate Header TXT and BIN (Merge/Only Structure!), 4=Generate TXT with DATA (Diff), 8=Export and Import DATA (Merge/Structure & Data)
*DBF_Conversion_Included: * && If DBF_Conversion_Support:4, you can specify multiple filemasks: www,fb2p_free.dbf
*DBF_Conversion_Excluded: && If DBF_Conversion_Support:4, you can specify multiple filemasks: www,fb2p_free.dbf
*-- DBC options
*- File per DBC options (UseFilesPerDBC: 1)
*OldFilesPerDBC: 0 && 1=Turns the File per DBC options on, 0 uses the old UseClassPerFile etc settings.
* && Options below will only read if OldFilesPerDBC is set 1 before!
* && If OldFilesPerDBC is set 0 later, alle setting will be lost
*UseFilesPerDBC: 0 && 0=One database dc2 file, 1=Multiple file.*.*.dc2 files
* && 0 creates only a file.dc2 with all DBC (file) data
* && 1 creates a file.dc2 with DBC properties
* && and additional DBC files per DBC item (stored-proc, table, ..)
* && Note: recration only if RedirectFilePerDBCToMain is 1
*RedirectFilePerDBCToMain 0 && 0=Don't redirect to file.dc2, 1=Redirect to file.tx2 when selecting file.item.*.dc2
*ItemPerDBCCheck: 0 && 0=Don't check file.item.*.dc2 inclusion, 1=Check file.item.*.dc2 inclusion
*DBF_BinChar_Base64: 1 && 0=For character type fields, if NoCPTrans 0=do not transform, 1=use Base64 transform (default)
*DBF_IncludeDeleted: 0 && 0=Do not include deleted records (default), 1=Include deleted records
*-- Class per file options (UseClassPerFile: 1)
*UseClassPerFile: 0 && 0=One library tx2 file, 1=Multiple file.class.tx2 files, 2=Multiple file.baseclass.class.tx2 files
*RedirectClassPerFileToMain: 0 && 0=Don't redirect to file.tx2, 1=Redirect to file.tx2 when selecting file.class.tx2
*ClassPerFileCheck: 0 && 0=Don't check file.class.tx2 inclusion, 1=Check file.class.tx2 inclusion
*-- Example configuration for SourceSafe compatibility:
*extension: pj2=pja
*extension: vc2=vca
*extension: sc2=sca
*extension: fr2=fra
*extension: lb2=lba
*extension: mn2=mna
*extension: db2=dba
*extension: dc2=dca
*-- Additional extensions
*extension: fk2=fkx && FKY
*extension: me2=fkx && MEM
*Settings changed
*----------------
UseClassPerFile: 2 && 0=One library tx2 file, 1=Multiple file.class.tx2 files, 2=Multiple file.baseclass.class.tx2 files
RedirectClassPerFileToMain: 1 && 0=Don't redirect to file.tx2, 1=Redirect to file.tx2 when selecting file.class.tx2
DBF_Conversion_Support: 0 && 0=No support, 1=Generate Header TXT only (Diff), 2=Generate Header TXT and BIN (Merge/Only Structure!), 4=Generate TXT with DATA (Diff), 8=Export and Import DATA (Merge/Structure & Data)
DBC_Conversion_Support: 0 && 0=No support, 1=Generate TXT only (Diff), 2=Generate TXT and BIN (Merge)
All use off this are the last four lines. But I like to keep the template (recent as version 1.20.0).
- Using single classes - because it's much more easy to merge a single class then a whole library.
- A single class selected might be added to the VCX
- Turn off DBC and DBF conversion
This project is part of VFPX.
Last changed: 2021/03/03 