Pommes

Pommes is a project checkout manager and database tool, you can use it search for projects and keep their checkouts organized on your disk. Here's an example:

~/Projects $ pommes goto lavender
[1]   Projects/svn.1and1.org/sales/tools/lavender
[2] A Projects/github.com/net/oneandone/lavender (git:ssh://git@github.com/mlhartme/lavender.git)
Please select: 2
[/Projects/github.com/net/oneandone] git clone ssh://git@github.com/mlhartme/lavender.git lavender
~/Projects/github.com/net/oneandone/lavender $ 

This command searches the database for lavender projects and lets you choose between the hits. Selecting 1 would simply cd into the already existing checkout, choosing 2 also creates the checkout.

Technically, Pommes is a command line tool that maintains a database with project metadata. Pommes can:

• index storages (e.g. github or github) to populate the database
• search the database by coordinates, dependencies, scm location, etc.
• perform (bulk-) checkouts for all projects that match a query.

Pommes detects addition metadata for Maven projects with a pom.xml file and php projects with a composer.json file.

The name Pommes stands for "many poms", "project mess", or a german word for french fries.

See also

Changes

Alternative tools

• https://github.com/gabrie30/ghorg
• https://github.com/x-motemen/ghq
• https://github.com/mthmulders/mcs

Setup

Prerequisites

• Linux or Mac
• Git
• Java 17 or higher
• Maven 3.6 or higher

Install the application

• open https://repo1.maven.org/maven2/net/oneandone/pommes/, choose the latest version and download the pommes-x.y.z-application.sh file, store it as executable pommes file in your path

    cd some-directory-in-your-path
    curl https://repo.maven.apache.org/maven2/net/oneandone/pommes/3.4.0/pommes-3.4.0-application.sh -o pommes
    chmod a+x pommes
  

• run pommes setup and follow the instructions. Basically, this sets up Pommes with root directory ~/Projects, configuration and other files will be stored in ~/Projects/.pommes. To set up a different directory, pass it as an argument on the command line

Managing the Database

Pommes maintains a Lucene database to store project metadata. The database is filled by indexing so-called storages. The database is stored in $POMMES_ROOT/.pommes/database, it's safe to delete this directory to wipe the database.

.pommes/config defines databases in the form storage. = {}

• name is an arbitrary name
• url is the primary definition of the storage, the protocol distinishes the storage type
  • github:
  • gitlab:
  • json:
  • file:///path-to-local-directory-with checkouts
• options depend on the storage type

Example urls

   storage.mlhartme=github:https://api.github.com %~mlhartme
   storage.company=gitlab:https://gitlab.company.com 

You can configure access token for github and gitlab by storing them in a file $POMMES_ROOT/.pommes/-.token. For example, a token in file $POMMES_ROOT/.pommes/github-mlhartme.token will be used to index the above mlhartme storage.

Find Command

Pommes stores the following fields for every project added to the database:

• id - zone and origin (i.e. url where the project descriptor (usually the pom) was loaded from)
• revision - last modified timestamp or content hash (to detect changes when re-adding a project)
• parent - parent pom coordinates; optional, may be missing
• artifact - coordinates of this project
• scm - where to find this project in a source code management system
• dep - coordinates of dependencies; optional list
• url - project url; optional, may be missing

You search the database with the find command: pommes find foo lists all projects that have a foo substring in their artifact or scm field. The default is to print the artifact field of matching projects. You can append -json to get all fields in json format instead.

Next, you can search for specific fields using one or multiple field identifiers. A field identifier is the first letter of the field name:

• pommes find d:bar lists projects with a bar substring in their dependencies
• pommes find dp:baz lists projects with a baz substring in their dependency or parent

(Technically, pommes find foo is a short-hand for pommes find :foo, and this in turn is a short-hand for pommes find as:foo)

You can also prepend fields with ! for negation. In this case, you should enclose the query argument in single quote, otherwise, the shell will expand the !. Example: pommes find '!d:bar' list all projects that have no dependency with a bar substring.

Next, you can combine queries with + to list projects matching both conditions. I.e. + means and.

• pommes find a:puc+d:httpcore lists projects with a puc substring in its artifact and httpcore in its dependencies.

Finally, you can combine queries with blanks to list projects matching one of the conditions. I.e. blank means or.

• pommes find d:puc d:pommes lists projects depending on either puc or pommes.

See the usage message below for the full query syntax.

TODO: Prefix, suffix, macros; formats

Checkout commands

Checkout commands (checkout, goto and st) manage checkouts on your disk, e.g. run scm operations on projects that match a query. Pommes supports Subversion and Git scms. The root directory for all checkouts is the directory containing .pommes.

Checkout

pommes checkout s:/mlhartme checks out all mlhartme projects.

Before changing anything on your disk, checkout presents a selection of the checkouts to perform, and you can pick one, multiple (separated with blanks) or all of them. Or you can quit without doing anything.

If a checkout already exists on your disk, it is not touched. This is useful to checkout newly created projects by just re-running your checkout command.

Goto

pommes goto sushi

searches the database for sushi, offers a selection of matching projects, checks it out if necessary, and cds into the resulting directory.

Status

pommes st lists all checkouts in the current directory, together with a status marker similar to git status.

Glossary for reading the source code

Descriptor: references a project, various implementation to load Maven, Json, etc

Project: holds project metadata: gav, parent, dependencies, scm, url

Storage: can be scanned for descriptors

Database: stores projects (in Lucene)

Usage Message

Project checkout manager and database tool.

Usage: 'pommes' ['-v'|'-e'] command args*

commands 'find' ('-output' str)? '-fold'? query ('-' format* | '-json' | '-dump' | '-'MACRO)? print projects matching this query; append '-json' to print json, '-dump' to print json without formatting; format is a string with placeholders: %c is replace be the current checkout and %FIELD_ID is replaced by the respective field; placeholders can be followed by angle brackets to filter for the enclosed substring or variables; output is a file or URL to write results to, default is the console. 'checkout' query checkout matching projects; skips existing checkouts; asks before doing any checkout 'goto' query offer selection of matching projects, checks it out when necessary, and cds into the checkout directory 'st' root? lists all checkouts under the specified directory (default '.') along with a status: ' ' - checkout is fine '?' - checkout is not in database '!' - checkout in wrong directory '#' - error checking this checkout 'index' {storage} re-index the specified (default: all) storages. 'setup' ['-batch'] {name'='value} creates '.pommes' directory with initial configuration containing name/values as storages; indexes all storages to create initial database; '.pommes' is created in the directory specified by $POMMES_ROOT, default is ~/Pommes.

fields in the database: (field id is the first letter of the field name.) origin Where this project was loaded from. Used as unique identifier. : revision Last modified timestamp or content hash of this pom. Used to detect changes. parent Coordinates of the parent project. artifact Coordinates of this project. dep Coordinates of project dependencies. scm Scm location for this project. Where to get the sources. url Url for this project. Where to read about the project.

query syntax query = '@' MACRO | or or = (and (' ' and))? index? index = NUMBER and = term ('+' term) term = field | lucene field = '!'? (FIELD_ID* match)? STR ; match on one of the specified fields (or 'ao' if not specified) match = ':' | '^' | '%' | '=' ; substring, prefix, suffix or string match lucene = '§' STR ; STR in Lucene query Syntax: https://lucene.apache.org/core/6_0_1/queryparser/org/apache/lucene/queryparser/classic/QueryParser.html and STR may contain the following variables: {gav} = coordinates for current project {ga} = group and artifact of current project {scm} = scm location for current directory

environment: POMMES_ROOT directory for manged checkouts and '.pommes'

Home: https://github.com/mlhartme/pommes